a1da95ef44
- P800 INT8 cannot sustain >1 concurrent long-context request without OOM; set all scenarios to concurrency=1. - Keep max_running=2 because max_running=1 triggers req_to_token_pool allocation failures in this image. - Reduce 64k max_context_len to 72000 (65536 + 4096 + pad). - Make run_bench.sh resilient to container death: detect dead container and skip remaining scenarios in the group instead of crashing the orchestrator. - Update README to document the concurrency limitation.
34 lines
926 B
Markdown
34 lines
926 B
Markdown
# P800 Long-Context Matrix for DeepSeek-V4-Flash-INT8
|
|
|
|
Long-context throughput/latency matrix for SGLang-XPU on Kunlun P800 (8x XPU).
|
|
|
|
## Scope
|
|
|
|
- Model: `DeepSeek-V4-Flash-INT8` (W8A8 INT8)
|
|
- Backend/engine: `sglang-xpu`
|
|
- Hardware: 8x Kunlun P800 XPU
|
|
- Input lengths: 64k, 128k
|
|
- Output lengths: 256, 1024, 4096
|
|
- Concurrency: **1** — P800 INT8 can only sustain one concurrent long-context request without OOM; `max_running=2` is kept because `max_running=1` triggers `req_to_token_pool` allocation failures in this image.
|
|
|
|
## Usage
|
|
|
|
```bash
|
|
cd /data1/yy/sskj/experiments/dsv4_p800_long_context_matrix
|
|
bash run_bench.sh
|
|
```
|
|
|
|
Optional environment overrides:
|
|
|
|
```bash
|
|
PORT=30013 RUN_ID=myrun bash run_bench.sh
|
|
```
|
|
|
|
## Outputs
|
|
|
|
Results are saved under `results/${RUN_ID}/`:
|
|
|
|
- `raw_outputs/`: raw `sglang.bench_serving` JSONL files
|
|
- `report.md`: parsed per-scenario metrics
|
|
- `results.json`: structured metadata + scenarios
|